JehanneOS/newlib
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

New expf, exp2f, logf, log2f and powf implementations

Browse Source 
Based on code from https://github.com/ARM-software/optimized-routines/

This patch adds a highly optimized generic implementation of expf,
exp2f, logf, log2f and powf.  The new functions are not only
faster (6x for powf!), but are also smaller and more accurate.
In order to achieve this, the algorithm uses double precision
arithmetic for accuracy, avoids divisions and uses small table
lookups to minimize the polynomials.  Special cases are handled
inline to avoid the unnecessary overhead of wrapper functions and
set errno to POSIX requirements.

The new functions are added under newlib/libm/common, but the old
implementations are kept (in newlib/libm/math) for non-IEEE or
pre-C99 systems.  Targets can enable the new math code by defining
__OBSOLETE_MATH_DEFAULT to 0 in newlib/libc/include/machine/ieeefp.h,
users can override the default by defining __OBSOLETE_MATH.
Currently the new code is enabled for AArch64 and AArch32 with VFP.
Targets with a single precision FPU may still prefer the old
implementation.

libm.a size changes:
arm: -1692
arm/thumb/v7-a/nofp: -878
arm/thumb/v7-a+fp/hard: -864
arm/thumb/v7-a+fp/softfp: -908
aarch64: -1476
This commit is contained in:
Szabolcs Nagy
2017-05-25 16:41:38 +01:00
committed by Corinna Vinschen
parent c165a27c01
commit c156098271
24 changed files with 1337 additions and 86 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
newlib/libc/include/machine/ieeefp.h
View File

@ -48,6 +48,23 @@
This represents what type a float arg is passed as. It is used when the type is
not promoted to double.
__OBSOLETE_MATH_DEFAULT
Default value for __OBSOLETE_MATH if that's not set by the user.
It should be set here based on predefined feature macros.
__OBSOLETE_MATH
If set to 1 then some new math code will be disabled and older libm
code will be used instead. This is necessary because the new math
code does not support all targets, it assumes that the toolchain has
ISO C99 support (hexfloat literals, standard fenv semantics), the
target has IEEE-754 conforming binary32 float and binary64 double
(not mixed endian) representation, standard SNaN representation,
double and single precision arithmetics has similar latency and it
has no legacy SVID matherr support, only POSIX errno and fenv
exception based error handling.
*/
#if (defined(__arm__) || defined(__thumb__)) && !defined(__MAVERICK__)
@ -61,6 +78,7 @@
# else
# define __IEEE_BIG_ENDIAN
# endif
# define __OBSOLETE_MATH_DEFAULT 0
#else
# define __IEEE_BIG_ENDIAN
# ifdef __ARMEL__
@ -75,6 +93,7 @@
#else
#define __IEEE_BIG_ENDIAN
#endif
#define __OBSOLETE_MATH_DEFAULT 0
#endif
#ifdef __epiphany__
@ -427,6 +446,14 @@
#define __IEEE_BIG_ENDIAN
#endif
#ifndef __OBSOLETE_MATH_DEFAULT
/* Use old math code by default. */
#define __OBSOLETE_MATH_DEFAULT 1
#endif
#ifndef __OBSOLETE_MATH
#define __OBSOLETE_MATH __OBSOLETE_MATH_DEFAULT
#endif
#ifndef __IEEE_BIG_ENDIAN
#ifndef __IEEE_LITTLE_ENDIAN
#error Endianess not declared!!